Message Passing Neural Networks

Also known as MPNN

Message passing neural networks study how to learn from graph-structured data by repeatedly sending, aggregating, and transforming information between connected nodes. Researchers vary the aggregation and update rules, incorporate edge and geometric features, and add mechanisms such as graph rewiring, virtual nodes, or global mixing to carry information farther. These models are used for node and graph prediction, molecular property estimation, recommendation, combinatorial optimization, and scientific simulation. A central tension is that local propagation is efficient and structure-aware but can lose long-range information or become indistinguishable across graphs; improving reach and expressivity without sacrificing efficiency or stability remains an open challenge.

Latest papers

64 papers8 papers in the last four weeks, up 167% on the four weeks before. 0.0% of all new papers.

Learn When and Where to Connect: Adaptive Virtual Nodes for Dynamic Message Passing on Graphs
Jun 2, 2026Jaejun Lee, Joyce Jiyoung WhangGraph Structure Learning Graph Neural Networks
School of Computing KAIST Daejeon, Republic of Korea · Department of AI Computing KAIST Daejeon, Republic of Korea
What drives performance in molecular MPNNs? An operator-level factorial benchmark
May 28, 2026Panyu Jiao, Shuizhou Chen, Yiheng Shen +3Message Passing Neural Networks Molecular Property Prediction
Materials Genome Institute, Shanghai University, Shanghai 200444, China · School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China · School of Materials Science and Engineering, Tongji University, Shanghai 201804, China
Invariant-Based Weight Sharing for Message Passing
May 25, 2026Florian SeiffarthGraph Neural Networks GNN Expressivity
University of Bonn, Bonn, Germany · Lamarr Institute for Machine Learning and Artificial Intelligence, Bonn, Germany
Evaluating passing decision-making in professional football: An enhanced MPNN approach to Receiver Selection
May 25, 2026Gabriel Masella, Giuseppe Alessio D'Inverno, Max Goldsmith +1Graph Neural Networks Message Passing Neural Networks
Department of Mathematics, Informatics and Geoscience, University of Trieste, via Weiss 2, Trieste, 34128, Italy. · MathLab, International School for Advanced Studies (SISSA), Via Bonomea 265, Trieste, 34136, Italy. · Royal Belgium Football Association, Rue de Bruxelles 480, Tubize, 1480, Belgium.
S $^3$ GNN: Efficient Global Mixing and Local Message Passing for Long-Range Graph Learning
May 22, 2026Dai Shi, Luke Thompson, Linhan Luo +4Long-Range Graph Learning Graph Neural Networks
Department of Engineering, University of Cambridge, Cambridge, UK · University of Sydney, Australia
The Cognitive Categorical Transformer: Category-Theoretic Inductive Biases for Language Modeling
May 22, 2026Al KariDecoder-Only Language Models Message Passing Neural Networks
Manceps Inc.
Learning over Positive and Negative Edges with Contrastive Message Passing
May 18, 2026Peter Pao-Huang, Charilaos I. Kanatsoulis, Michael Bereket +1Graph Neural Networks Graph Representation Learning
Department of Computer Science Stanford University · Stanford University
Graph Neural Networks with Triangle-Based Messages for the Multicut Problem
May 13, 2026Jannik Irmai, Lucas Fabian Naumann, Bjoern AndresGraph Neural Networks Combinatorial Optimization
Faculty of Computer Science, TU Dresden · Center for Scalable Data Analytics and AI, Dresden/Leipzig
MLGIB: Multi-Label Graph Information Bottleneck for Expressive and Robust Message Passing
May 13, 2026Chaokai Wu, Haofu Shi, Ningxuan Ma +2Multi-Label Classification Graph Neural Networks
From Message-Passing to Linearized Graph Sequence Models
May 12, 2026Joël Mathys, Basil Rohner, Saku Peltonen +1Long-Range Graph Learning Graph Representation Learning
1ETH Zurich, Switzerland
Debiasing Message Passing to Mitigate Popularity Bias in GNN-based Collaborative Filtering
May 11, 2026Md Aminul Islam, Ahmed Sayeed Faruk, Sourav Medya +1Graph Neural Networks Popularity Bias
University of Illinois Chicago
Scalable Mamba-Based Message-Passing Neural Decoder for Error-Correcting Codes
May 11, 2026Rostislav Gusev, Nikita Aleksandrov, Artem Solomkin +1Neural Decoding Message Passing Neural Networks
Skolkovo Institute of Science and Technology, Moscow, Russia
The Polynomial Counting Capabilities of Message Passing Neural Networks
May 11, 2026Marco Sälzer, Pascal Bergsträßer, Anthony W. LinGraph Neural Networks Message Passing Neural Networks
1RPTU University Kaiserslautern-Landau · 2Max Planck Institute for Software Systems (MPI-SWS)
No Triangulation Without Representation: Generalization in Topological Deep Learning
May 7, 2026Johannes S. Schmidt, Martin Carrasco, Ernst Röell +3Neural Network Generalization Graph Neural Networks
Dept. of Math. & Statistics University of Montréal Montréal, Canada · Dept. of Informatics University of Fribourg Fribourg, Switzerland
Geometry-Aware Simplicial Message Passing
May 7, 2026Elena Xinyi Wang, Bastian RieckMessage Passing Neural Networks Geometric Representation Learning
AIDOS Lab, University of Fribourg Fribourg, Switzerland
Graph Neural Networks in the Wilson Loop Representation of Abelian Lattice Gauge Theories
May 5, 2026Ali Rayat, Gia-Wei ChernNeural Surrogate Modeling Graph Neural Networks
Department of Physics, University of Virginia, Charlottesville, VA 22904, USA
A Functorial Formulation of Neighborhood Aggregating Deep Learning
Apr 27, 2026Sun Woo Park, Yun Young Choi, U Jin Choi +1Message Passing Neural Networks Convolutional Neural Networks
Towards Understanding the Expressive Power of GNNs with Global Readout
Apr 23, 2026Maurice Funk, Daumantas KojelisGraph Neural Networks GNN Expressivity
Leipzig University, Germany
Gauge-Equivariant Graph Neural Networks for Lattice Gauge Theories
Apr 22, 2026Ali Rayat, Yaohang Li, Gia-Wei ChernMessage Passing Neural Networks High-Energy Physics
Department of Physics, University of Virginia, Charlottesville, VA 22904, USA · Department of Computer Science, Old Dominion University, Norfolk, VA 22904, USA.
Concept Graph Convolutions: Message Passing in the Concept Space
Apr 22, 2026Lucie Charlotte Magister, Pietro LioGraph Neural Networks Message Passing Neural Networks
University of Cambridge
RoleMAG: Learning Neighbor Roles in Multimodal Graphs
Apr 14, 2026Yilong Zuo, Xunkai Li, Zhihan Zhang +2Multimodal Graph Learning Graph Neural Networks
Beijing Institute of Technology, Beijing, China
Hypergraph Neural Diffusion: A PDE-Inspired Framework for Hypergraph Message Passing
Apr 13, 2026Zhiheng Zhou, Mengyao Zhou, Xixun Lin +2Hypergraph Neural Networks Hypergraph Learning
School of Mathematics and Statistics, Shandong University · Academy of Mathematics and Systems Science, Chinese Academy of Sciences · School of Mathematical Sciences, University of Chinese Academy of Sciences +2
A Tight Expressivity Hierarchy for GNN-Based Entity Resolution in Master Data Management
Mar 28, 2026Ashwin GanesanEntity Resolution GNN Expressivity
Centre for Cybersecurity, Trust and Reliability (CyStar), Department of Computer Science and Engineering, Indian Institute of Technology, Madras, Tamilnadu, India.
LLM as Graph Kernel: Rethinking Message Passing on Text-Rich Graphs
Mar 16, 2026Ying Zhang, Hang Yu, Haipeng Zhang +1LLMs for Graph Learning Text-Attributed Graph Learning
Ant Group · ShanghaiTech University
Lost in Aggregation: On a Fundamental Expressivity Limit of Message-Passing Graph Neural Networks
Mar 16, 2026Eran RosenbluthGraph Neural Networks GNN Expressivity
RWTH Aachen University
$P^2$ GNN: Two Prototype Sets to boost GNN Performance
Mar 10, 2026Arihant Jain, Gundeep Arora, Anoop Saladi +1Graph Neural Networks Message Passing Neural Networks
Amazon, India
Local Message-Passing for Discrete Graph Generation
Mar 9, 2026Jay Revolinsky, Harry Shomer, Jiliang TangGraph Neural Networks Graph Generation
Michigan State University East Lansing, Michigan, USA · University of Texas - Arlington Arlington, Texas, USA
Training-Free Cross-Architecture Merging for Graph Neural Networks
Feb 22, 2026Rishabh Bhattacharya, Vikaskumar Kalsariya, Naresh ManwaniModel Merging Graph Neural Networks
International Institute of Information Technology, Hyderabad, India
Learning to Approximate Uniform Facility Location via Graph Neural Networks
Feb 13, 2026Chendi Qian, Christopher Morris, Stefanie Jegelka +1Graph Neural Networks Combinatorial Optimization
RWTH Aachen University, Germany · Technical University of Munich, Germany · Massachusetts Institute of Technology, USA +1
Enhancing Graph Representations with Neighborhood-Contextualized Message-Passing
Nov 14, 2025Brian Godwin Lim, Galvin Brice Lim, Renzo Roel Tan +2Graph Neural Networks Graph Representation Learning
Nara Institute of Science and Technology, Kyoto University, and Ateneo de Manila University · UNI-President Information Philippines Corporation · Nara Institute of Science and Technology and Ateneo de Manila University +2

Jun 2, 2026J/K move · Enter open · S save

Learn When and Where to Connect: Adaptive Virtual Nodes for Dynamic Message Passing on Graphs

Jaejun Lee, Joyce Jiyoung Whang

School of Computing KAIST Daejeon, Republic of Korea · Department of AI Computing KAIST Daejeon, Republic of Korea

Graph Structure Learning Graph Neural Networks Graph Representation Learning Message Passing Neural Networks

Open paper AskPDF arXiv

While Virtual Nodes (VNs) are often utilized in Message Passing Neural Networks (MPNNs) to facilitate effective message passing, existing VN-based methods have limitations, such as constraining all nodes to connect to the same number of VNs, fixing the connections before applying MPNNs, and connecting a node to a VN independently of the other nodes that connect to the same VN. We propose MAVN, an end-to-end differentiable MPNN framework that allows non-constrained connections between nodes and VNs and dynamically introduces VNs on demand in response to evolving node representations across layers. Specifically, MAVN learns to adaptively determine when (at which layer) and where (to which nodes) to introduce and connect VNs based on the relative importance of connections. From a pool of candidate VNs, MAVN selects the necessary VNs in each layer, where each selected VN is connected to a nonempty subset of nodes, guided by a dual-perspective scoring mechanism that jointly captures the nodes' preferences for VNs and the VNs' preferences for nodes. We theoretically prove that for any node-VN connectivity pattern, there exists a set of MAVN's parameters that can simulate the pattern. Experiments on nine real-world datasets demonstrate that MAVN consistently improves the performance of backbone MPNNs, achieving up to 46.5% improvement over the backbones and outperforms the baselines.